Le moteur NLP d'Oupi scanne votre message en temps réel et identifie automatiquement toutes les données sensibles : noms, emails, téléphones, IBAN, adresses, numéros de dossier, montants. Aucune intervention manuelle requise.
Chaque donnée sensible est remplacée par un token unique ([PERSONNE_1], [EMAIL_1]...) avant d'être envoyée au modèle IA. Le modèle ne voit jamais vos vraies données — il travaille avec des références anonymisées.
La réponse de l'IA est automatiquement dé-tokenisée : les tokens sont replacés par vos vraies données. Vous recevez un texte complet, prêt à l'emploi, sans aucune manipulation.
Most AI platforms send your data as-is to models. Oupi is the only platform that intercepts, anonymizes and restores sensitive data transparently — without impacting response quality. No other solution combines real-time tokenization, multi-tool compatibility and native GDPR compliance in a single system.
🔐 Oupi vs. other platforms
❌ ChatGPT, Gemini, etc.
✓ Oupi + Privacy Protection
🧠 Smart detection — Recognized types
Jean Dupont → [PERSON_1]
[email protected] → [EMAIL_1]
+33 6 12 34 56 → [PHONE_1]
FR76 3000... → [IBAN_1]
12 rue de Paris → [ADDRESS_1]
€3,450 → [AMOUNT_1]
#F-2026-0847 → [REF_1]
Acme Corp → [COMPANY_1]
1 85 12 75... → [ID_1]
4532 •••• → [CARD_1]
Oupi's specialized NLP engine automatically identifies sensitive data in your texts, without you needing to flag them. Names, emails, phones, IBANs, addresses, credit card numbers, amounts, references — everything is detected and tokenized in real time before sending to AI.
Privacy Protection is not an isolated module. It's natively integrated into every Oupi feature — from Chat to Studio to Workflows and Bots.
Every message sent in Chat is automatically scanned and tokenized. You chat normally — the protection is invisible and total.
🎨Create documents, presentations, emails with client data. Privacy Protection protects everything before submitting it to the AI model.
⚙️Automate tasks involving sensitive data (client follow-ups, HR reports, invoices). Tokens circulate instead of your real data.
🤖Your chatbots deployed on website or WhatsApp automatically protect visitor data. GDPR compliant by default.
🎥Meeting transcriptions containing names, contacts and figures are tokenized before AI analysis. Reports remain complete.
📈Client data used for content personalization remains anonymized during AI processing.
Privacy Protection was designed from the start for European compliance. Tokens are stored locally on our European servers, never transmitted to AI providers. A complete audit log allows you to trace every tokenization.
📋 Audit log — Privacy Protection
Tokens and data stored exclusively on servers in Europe. GDPR compliance, no transfer outside EU.
All token ↔ real data mappings are encrypted with AES-256. Even in case of breach, data remains unreadable.
Delete your tokens and associated data at any time. No hidden retention, full control.
Only authorized users can view token/data mappings. Granular permissions per team.
Exportez des rapports d'audit pour vos DPO et audits RGPD. Traçabilité complète de chaque tokenisation.
Set token retention duration: 24h, 7 days, 30 days or custom according to your internal policies.